I am trying to wrap my mind around the following scenario.
A slurry, SG=1.05 @ 160F and 3000USgpm, runs through an 18" header. From there it ties into a "distribution box" where the slurry is spread to a cluster of cyclones with a springloaded valve between each cyclone and box. The overflow of each cyclone goes into tank1, the underflow into tank2. Some distance upstream of the distribution box a 10" bypass lines tees of the 18" header and connects to tank2 (from there the underflow slurry is redirected to a different process). The process is continous.
The distribution box has a MAWP of around 170PSIg, the operating pressure is around 70PSIg. I want to put a rupture disk in the bypass line to protect the distribution box. Let's assume for now it is set to burst at 150PSIg (taking into account backpressure, manufacturing tolerance, burst tolerance etc).
The main failure scenario is a power failure which shuts the valves close. Since the valves shut immediately, I need to check whether the pressure rise in that case is high enough to burst the rupture disk.
1) Can I assume that there is a valve just downstream of the box that closes shut immediately, instead of having to deal with the geometry of the box (spherical) and having to ponder whether some pressure waves do get reflected and interfer with each other etcetc?
2) As the pressure wave travels downstream the header, and reaches the T, two things should (will?) happen. The pressure wave continues going down the main header (until it hits a pump a few hundred meters downstream), and the rupture disks bursts. When the rupture disks bursts the slurry should leave through the bypass line (given the line size is correct to relief the required flow).
So, essentially, a pressure wave goes down the main header, "telling" the slurry about a zero-velocity condition close to the distribution box. However, as that wave is travelling, the rupture disk in the bypass line bursts and the fluid upstream of the pressure wave starts flowing again? Is there some other pressure wave travelling downstream the main header when the rupture disk bursts?
I think, although the process setup looks realatively simple, the transient scenario is quite intricate. I am curious to hear what thoughts others might have about this scenario and whether I am simplifying things too much.
The attached drawing is simplified a lot (no valves shown between box and cyclones etc).